>>For processes that could possibly timeout it's best to move those off to an asynchronous process. WWWC comes with a class for doing just that--lookup the wwAsyncWebRequest class in the help file. If you have the WebRAD book, there's also a clear example of how an async request works.
>
>I've been thinking about that approach. However, with a better infrastructure, I need to get those results immediately. So, I'll probably give a try with SQL Server for the search engine as I could use the builtin mechanisms. If it doesn't have any bug with large table being indexed, then that could resolve the issue.

I doubt that going to SQL Server would improve the performance. You could try it. At least it would mean that you're not relying on a discountinued product for your search. IAC, you should still do this asynchronously because there will always be some kind of search request that will cause a timeout. Keep in mind that even with an async request, if the results come back right away, you can send the results to the client right away. It just takes one more hit than it would otherwise.